In a 2018 article entitled “Human Cognitive Limitations”, the University of Utah’s Alan Morris estimated the number of variables an intensivist has to take into account for a patient on mechanical ventilation.

He came up with 236. A typical ward round for Brijesh Patel, an intensivist at the Royal Brompton hospital in London, involves checking the settings on 20–30 ventilators.

It is an enormously time-consuming process. “The question of how to manage the ventilator is at the centre of critical care”, points out Patel. “Yet we are typically working without a deep knowledge of what is happening with patients on a breath-by-breath basis, or from hour to hour. We have to find a better way to do things.”

Critical care is a specialty for which artificial intelligence (AI) holds particular promise. After all, the intensive care unit (ICU) is a space in which vast amounts of data are routinely collected, making it ideal for deploying machine-learning techniques.

“There are plenty of areas in critical care where it would be extremely helpful to have efficacious, fair, and transparent AI systems”, notes Gary Weissman, assistant professor in pulmonary and critical care medicine at the University of Pennsylvania Perelman School of Medicine.

Read More